DRAFT Reliability Guideline: Modeling Distributed Energy Resources in Dynamic Load Models

Transcription

1 DRAFT Reliability Guideline: Modeling Distributed Energy Resources in Dynamic Load Models Problem Statement With the proliferation of distributed energy resources (DER), Transmission Planners must adapt models and modeling practices to account for and differentiate end use loads and the offset in net demand by these resources. In the past, and at lower penetrations of DER integrating into the distribution system, net load reduction has been used. This assumes the same load composition for these different demand levels; however, it is well understood and expected that net load reduction is actually a result of the same or greater demand with an offset due to some level of distributed energy resources. However, these practices are not sustainable moving forward as the distribution system continues to integrate more DER. The NERC Distributed Energy Resources Task Force (DERTF) is developing a report 1 that will provide definitions for DER including different types of DER such as distributed generation (DG), behind the meter generation (BTMG), demand-side management (DSM), and others. These definitions will be used as the prevailing definition of DER; however, for the purposes of modeling DER and the guidance provided herein, a simplified set of definitions are used. The recommended modeling practices differentiate between types of generating resources by the location of their interconnection to the distribution system. For the purposes of dynamic load modeling specified in this guidance, the following definitions are used: Utility-Scale Distributed Energy Resources (U-DER): distributed energy resources directly connected to the distribution bus 2 or connected to the distribution bus through a dedicated, nonload serving feeder. These resources are specifically three-phase interconnections, and can range in capacity, for example, from 0.5 to 20 MW although facility ratings can differ. Retail-Scale Distributed Energy Resources (R-DER): distributed energy resources that offset customer load. These DER include residential 3, commercial, and industrial customers. Typically, the residential units are single-phase while the commercial and industrial units can be single- or threephase facilities. 1 The DERTF Final Report will be completed by end of year 2016 and available here. 2 The distribution bus is connected to a transmission voltage bus via the transmission-distribution transformer. Resources not directly connected to this bus do not meet the criteria for this definition. 3 This also applies to community distributed energy resources that do not serve any load directly but are interconnected directly to a distribution load serving feeder.

2 Discussion of Modeling Practices This section includes a description of modeling practices in both the powerflow and dynamics base case and how utility-scale distributed energy resources (U-DER) and retail-scale distributed energy resources (R- DER) can be represented to effectively capture their performance. Simple Powerflow Modeling Aggregated loads are commonly modeled in the powerflow base cases at the high side of the distribution transformer, with no representation of the transmission-distribution transformer that steps the voltage down to feeder levels (Figure 1). This is generally done to simplify modeling for studying the impacts on Bulk-Power System (BPS) performance. Figure 1: Powerflow Representation of Aggregated Loads Modeled Distribution Transformer One form of modeling loads is to represent each transmission-distribution transformer as a point of connectivity and represent that in the powerflow base case. In this situation, the distribution transformer is modeled explicitly, capturing the electrical impedance as well as any under-load tap changing (ULTC) impacts the transformer may have in steady-state pre- and post-contingency operating conditions. This is particularly important for voltage stability analysis where the transformer impedance and ULTC operation can impact longer term dynamics. Figure 2 shows an example of this representation. Figure 2: Transmission-Distribution Transformer Representation Modeling Guidance: DER in Dynamic Load Models 2

3 Composite Load Model Representation in Stability Programs The most detailed representation of end-use load for dynamic simulations and transient stability assessment available in commercial software platforms is the Composite Load Model (CLM). This model is used in this guidance document to explain how U-DER and R-DER can be incorporated into the model and modeling practices. The CLM includes different types of induction motor models, electronic load, and static load. It also includes representation of the transmission-distribution transformer, shunt compensation at the distribution level, and a distribution feeder equivalent impedance. Lastly, protective tripping can be modeled for each machine as well as the ULTC action. The CLM in the form shown in Figure 3 does not allow for representation of U-DER or R-DER. This is the current implementation of the CLM model structure in at least four major commercial software platforms in North America. The model is not well suited for breaking out U-DER or R-DER and separating the amount of net load reduction (U-DER and R-DER) versus gross load modeled as various forms of induction motors and other types of load. Figure 3: Composite Load Model DER in the Composite Load Model One option is to include all DER directly into the dynamic load model, as shown in Figure 4. This capability is included in a modified version of the CLM presently in one commercial software platform 4. In this case, the DER model represents an aggregate of all U-DER and R-DER on the distribution feeder and is located at the load-bus end of the feeder. This is a simplification to represent some average impedance to all the loads (and DER) on the feeder. While this may be a reasonable modeling assumption at low penetration levels, the U-DER and R-DER must therefore be represented by a uniform control strategy ( DER model). In this case, the powerflow base cases load records must have fields to account for the amount of DER. Increasing DER will result in net load reduction; however, the gross load will be broken out in the dynamic 4 cmpldw in GE PSLF TM. Modeling Guidance: DER in Dynamic Load Models 3

4 simulations such that a more accurate amount of induction motor load and other loads will be represented as well as a closer representation of DER transient performance. Again, the latest versions of several commercial software platforms used in North America presently have this feature of separate fields in the powerflow load records for specifying active and reactive power associated with aggregated DER for each load modeled in the powerflow. Figure 4: Inclusion of DER in the Composite Load Model U-DER and R-DER in the Composite Load Model Another option is to include U-DER and R-DER in the dynamic load model, as shown in Figure 5. In this case, the U-DER and R-DER are separated into different aggregate models with R-DER represented at the load bus of the feeder model and U-DER represented at the distribution substation. The benefits of this approach include: Effects of locationality of U-DER and R-DER: The equivalent feeder impedance results in several percent (e.g., typically around 5%) voltage drop between the low-side bus and load bus. During transient conditions, that voltage difference may cause some DER to trip offline while others may ride through. Control strategies: Interconnection requirements for U-DER may differ from R-DER. Larger-scale installations (U-DER) may have more complex controls such as low voltage ride through, voltage control, and other plant level controls. On the other hand, small-scale (R-DER) likely have simpler control features (e.g., present installations likely adhere to the original IEEE 1547 requirements, which were essentially constant power factor control with no ride through capability). Modeling Guidance: DER in Dynamic Load Models 4

5 The drawbacks of this approach are: Complicated Data Entry: The powerflow base case would need to include the U-DER and R-DER separately or a fractional differentiation would have to be accounted for in the dynamic load model. This would result in complications in data management and for running sensitivity studies. Dynamic Load Model Complexity: In addition to the complexity introduced in the powerflow solution, the composite load model records would also become increasingly complicated by multiple types of U-DER and R-DER. Figure 5: Inclusion of U-DER and R-DER in the Composite Load Model Recommended Modeling Practice for U-DER and R-DER in the Dynamic Load Models The NERC LMTF has discussed and explored options for effectively accounting for U-DER and R-DER in the dynamic simulations as well as how to account for these resources in the powerflow base case as well. Both are interlinked with one another so there is no realistic way to separate the dynamics and powerflow cases and how these resources are modeled. It is recommended that Transmission Planners and Planning Coordinators, in conjunction with the Distribution Providers, identify U-DER and R-DER thresholds for which these resources must be explicitly modeled in the powerflow and dynamics cases as distinct resources. The TPs and PCs should clearly differentiate in the model the resources that are U-DER and the resources that are distributed R-DER. This will clearly differentiate how these resources are modeled in the dynamic simulations as well as in the powerflow base case for contingency analysis and sensitivity analysis. The thresholds, for example, should be based on an individual resource s impact on the system as well as an aggregate impact. Modeling Guidance: DER in Dynamic Load Models 5

6 Gross aggregate nameplate rating of an individual U-DER facility directly connected to the distribution bus or interconnected to the distribution bus through a dedicated, non-load serving feeder; and Gross aggregate nameplate rating of all connected R-DER resources that offset customer load including residential, commercial, and industrial customers. Table 1 shows an example set of requirements for modeling U-DER and R-DER, with thresholds determined based on engineering judgment applicable to the TP or PC electrical characteristics and processes. U-DER Modeling: Any individual U-DER facility rated 5 MVA or higher should be modeled explicitly in the powerflow case at the low-side of the transmission-distribution transformer. A dynamics record, such as PV1 6 model, is required to account for the transient behavior 7 of this plant. R-DER Modeling: If the gross aggregate nameplate rating of R-DER exceeds this threshold, these distributed resource should be accounted for in dynamic simulations as part of the dynamic load model. While this may not require any explicit representation in the powerflow base case, the amount of R-DER should be accounted for in the load record and/or integrated into the dynamic model. Table 1: Example of U-DER and R-DER Requirements Criteria Description Threshold U-DER Modeling R-DER Modeling Gross aggregate nameplate rating of an individual U-DER facility directly connected to the distribution bus or interconnected to the distribution bus through a dedicated, non-load serving feeder Gross aggregate nameplate rating of all connected R-DER resources that offset customer load including residential, commercial, and industrial customers MVA MVA Figure 6 shows the conventional powerflow representation of the load in a powerflow base case and the recommended representation that explicitly models U-DER above a given size threshold. Note that each U- DER above the threshold would be modeled explicitly via its own step-up transformer, as applicable, to the low-side bus. The load is also connected to the low-side bus. 5 This is intentionally left blank as a template or placeholder for applying this in a particular TP or PC footprint. 6 The PV1 is a generic dynamic model for a photovoltaic inverter system. 7 Depending on complexity of the actual U-DER, for inverter coupled U-DER, more sophisticated models such as the second generation generic renewable energy system models may also be used (i.e. regc_a, reec_b and repc_a). Other U-DER (e.g. combined heat and power using conventional gas and steam-turbine generators) can also be modeled using standard models available in commercial software platforms. Modeling Guidance: DER in Dynamic Load Models 6

7 Figure 6: Representing Utility-Scale DER (U-DER) in the Powerflow Base Case Once represented in the powerflow model in this manner, the data for the composite load model should be modified to account for explicit representation of the U-DER and transmission-distribution transformer. Figure 7 shows the composite load model where the distribution transformer impedance is not represented in the dynamic record, it is modeled explicitly in the powerflow to accommodate one or more U-DER. The transformer impedance is not represented in the CLM (impedance set to zero in the dynamic load model); therefore, any LTC modeling would have to be done outside the CLM such as using the ltc1 model available 8. The motor load and distribution equivalent feeder impedance is modeled as part of the CLM, and the R- DER are represented at the load bus based on the input in the powerflow load record while the load is fully accounted for rather than any net load reduction. Figure 7: Dynamic Load Model Representation with U-DER Represented in the Powerflow Base Case 8 Software vendors are exploring the concept of applying an area-, zone-, or owner-based LTC model that could be applied to all applicable transformers to address LTC modeling. Modeling Guidance: DER in Dynamic Load Models 7

8 To capture the R-DER in the powerflow solution, the load records must have the capability 9 to input the R- DER quantity in the powerflow. Figure 8 shows an example of the R-DER included in the powerflow load records. The red box shows the R-DER specified, for example 80 MW and 20 Mvar of actual load with 40 MW and 0 Mvar of R-DER at Bus 2. The blue box shows the net load equal to the actual load less the R-DER quantity specified for MW and Mvar, defined as: NNNNNN MMMM = MMMM llllllll DDDDDDDD MMMM RR DDDDDD NNNNNN MMMMMMMM = MMMMMMMM llllllll DDDDDDDD MMMMMMMM RR DDDDDD Figure 8: Capturing R-DER in the Powerflow Load Records (Source: PowerWorld) The R-DER captured by these quantities in the powerflow would be based on the threshold values as specified in Table 1 for R-DER Modeling. 9 Some software platforms have already adopted this approach; the NERC LMTF recommends all software platforms adopt the same approach to unify this modeling practice and enable flexibility for capturing DER as part of the load records. Modeling Guidance: DER in Dynamic Load Models 8